Several experimental model systems were used to further test the hypothesis that glycogen synthase kinase-3 (GSK-3) and histone deacetylases (HDACs) are initial targets of lithium and valproic acid (VPA), respectively, that trigger diverse neurobiological effects. Our studies identified a number of novel signaling pathways and neuroprotective and neurotrophic genes targeted by lithium and VPA, and underlying mechanisms were characterized in both in vitro and in vivo experimental settings. Expanded studies on the effects of mood stabilizers in a rat ischemic stroke model, and mouse models of HD and stress-induced depression were performed. Results of these studies identified previously unknown neurobiological actions induced by these mood stabilizers. In primary brain neurons, we identified a previously unknown function of fibroblast growth factor-21 (FGF-21), namely its ability to mediate the synergistic neuroprotective effects of lithium and VPA against glutamate excitotoxicity (Leng et al., Mol Psychiatry, 2015). Until now, FGF-21 was thought to be expressed only in the peripheral systems, have a prominent role in regulating glucose and fatty acid metabolism, and act as a putative therapeutic target for diabetes and obesity. Our results demonstrated for the first time that FGF-21 can be markedly induced in primary brain neurons and intact brain of rodents following co-treatment with lithium and VPA. More importantly, FGF-21 mediated, at least in part, the synergistic neuroprotection induced by lithium-VPA treatment. FGF-21s neuroprotection involved Akt-1 activation and GSK-3 inhibition; interestingly, both events also reciprocally regulated FGF-21 induction. Our recent results showed that FGF-21 RNA and protein levels were robustly decreased in the brains of ischemic rats (Wang, Z et al., submitted) and ALS mice (Wang J et al., in preparation), suggesting that this growth factor is a novel target for therapeutic intervention of brain disorders. In a very recent study, we found that rat C6 glioma cells treated with VPA had an increase in mRNA levels of FGF-21, the length of neurite-like processes and histone acetylation level, a marker for HDAC inhibition. Inhibition of HDAC 2 or 3 but not HDAC1 isoform markedly increased FGF-21 mRNA levels. In primary rat cortical glial cells, VPA, SB, SAHA and TSA also significantly increased FGF-21 mRNA levels, which were associated with elongated length of astrocyte processes (Leng et al., submitted). Continuing our studies using middle cerebral artery occlusion (MCAO) in rats as a model of focal cerebral ischemia, has led to a number of original findings. We reported that HDAC inhibition by a VPA analog, sodium butyrate, enhanced MCAO-induced oligodendrogenesis in multiple ischemic brain regions, and this effect required activation of BDNF-TrkB signaling (Kim et al., AJTR, 2014). Furthermore, MCAO rats treated with a specific HDAC6 isoform inhibitor, tubastatin A, showed remarkable neuroprotective effects and behavioral improvements with a time window of at least 24 hours after insult. The beneficial effects of tubastatin A were associated with tubulin hyperacetylation, and may involve neuroprotection against excitotoxicity and amelioration of defective mitochondrial transport via upregulation of FGF-21(Wang, Z et al., submitted). We also studied the regulation of miRNAs by mood stabilizers to identify novel miRNA-mediated signatures and mechanisms in lymphoblastoid cell lines (LCLs) derived from bipolar patients who were lithium responders or non-responders. Several prominent miRNAs notably Let-7 were identified by microarray to be differentially regulated in bipolar responders and non-responders, and their interactions with target mRNAs in LCLs have been studied by GRANITE (Hunsberger et al., Translational Psychiatry, 2015). In a pilot clinical study in collaboration with Dr. Giia-Sheun Peng in Taiwan, we found that three-month treatment with VPA commencing 3-24 hours after stroke markedly reduced neurological deficits (Lee et al., Am J Transl Res, 2014). In another clinical trial in collaboration with Dr. Hong Jiang of Xiangya Hospital in China, the safety and efficacy of VPA treatment in SCA3/MJD patients were tested. Both clinical measurements and motor function were found to be significantly improved after 12-week VPA treatment (Lei et al., submitted). In collaboration with same group, we found that miR-25 alleviated polyQ-mediated cytotoxicity by silencing ATXN3, suggesting another possible therapeutic intervention for SCA3/MJD (Huang et al., FEBS Lett, 2014). In conjunction with Dr. Honglin Feng of Harbin Medical University, we found that AEG-1 is critical for cell survival in ALS motor neurons in vivo and in vitro and that the defective PI3K/Akt/CREB/AEG-1 cycle is likely involved in the pathophysiology of ALS and the treatment target of lithium (Yin et al., Mol Cell Neurosci, in press). We assessed the therapeutic potential of dietary treatment with lithium and/or VPA in two transgenic mouse models of HD, N171-82Q and YAC128. We found that daily dietary co-treatment with lithium and VPA more effectively alleviated impaired locomotion and depressive-like behaviors than mono-treatment in both mouse models, and significantly prolonged the lifespan of N171-82Q mice (Chiu et al., Neuropsychopharmacol, 2011; Scheuing et al., Int J Biol Sci, 2014). Levels of BDNF in the brains of both strains were also more consistently elevated by lithium-VPA co-treatment. Recently, we observed that long-term administration of a BDNF TrkB receptor agonist, LM22A, elicited behavioral benefits in N171-82Q mice (Chiu et al., in preparation), further supporting the roles of BDNF/TrkB signaling in HD pathology and therapy. In another ongoing study using N171-82Q mice, we found that intranasal delivery of mouse mesenchymal stem cells (MSC) preconditioned with both lithium and VPA ameliorated behavioral deficits. Histological analysis confirmed that intranasally administered MSCs were detected in the brain, and behavioral tests showed significant improvements in motor function. Gene expression profiling of preconditioned MSCs revealed a robust increase in expression of genes involved in trophic effects, antioxidant, anti-apoptosis, cytokine/chemokine receptor, migration, mitochondrial energy metabolism, and stress response signaling pathways (Linares et al., submitted). These findings suggest a novel avenue for HD therapeutic intervention. An additional investigation using aged YAC128 HD mice found that lithium and VPA treatment can clear mutant Htt protein, the disease-causing protein, , as well as lead to behavioral improvements (Liao et al., in preparation). We have also completed a project related to the neurobiology of lithium: Pre- or post-treatment with lithium was shown to potentiate the rapid antidepressant effects of ketamine in a chronically stressed model of depression in mice (Chiu et al., IJNN, 2015; Scheuing et al., Front Neurosci, 2015). This potentiation by lithium involved a robust decrease in ketamine dose-requirement, prolongation of its antidepressant duration, and dendritic spine increase and suppression of ketamine-induced oxidative stress. Further, ketamine-induced activation of the BDNF and mTOR signaling pathways were synergistically enhanced by lithium-ketamine co-treatment. In summary, our recent work has markedly increased our understanding of the molecular and cellular actions of mood stabilizers, and substantially advanced our knowledge of their effects in a number of experimental models of neurodegenerative and neuropsychiatric diseases. With the completion of a number of ongoing projects, we expect to provide further mechanistic insights and set the stage for clinical investigations into the use of mood stabilizers to intervene in certain CNS disorders.